have a cloned git available and be able to run the build-scripts.

The system can always be reset to a clean state by using “deploy_full.sh”. Usually this is not needed. In case you may also deploy specific components by using “deploy_config_tool”, “deploy_controller” or “deploy_downloaders”.

If you dont know what component causes the trouble, start with the config tool and grep through the docker logs. Then go on to the controller.

=Downloaders= Use the table “exceptions” on the krawla-db to track down the workers that cause the trouble. Alternatively use the config-tool to vnc into some random workers. A worker that has a problem either hangs or restarts constantly. Use the ip and local hostname you got from the vnc or the table to ssh into the server that has the problem and check the logs of that specific service.